diff --git a/docs/scripts/blog-articles-manifest.json b/docs/scripts/blog-articles-manifest.json
index db0513f9a..252d9521e 100644
--- a/docs/scripts/blog-articles-manifest.json
+++ b/docs/scripts/blog-articles-manifest.json
@@ -445,6 +445,11 @@
       "sha256": "1e2ee21c697ab9c2320094874889e2ac0bd2fc44efb529eafa4d895393ca5691",
       "bytes": 161907
     },
+    {
+      "path": "/images/articles/langbot-wechat-official-account-qr.png",
+      "sha256": "233aede79ed71fd721f1f77c839d726b9dd8ed667f5e951d994c19f962c6bff5",
+      "bytes": 14931
+    },
     {
       "path": "/images/articles/source/images/article-mcp/01-init-page.png",
       "sha256": "b392e42d582c8c1723a2eb63596801b9d459df0465377246d9a4015edcae1ca0",
@@ -589,6 +594,31 @@
       "path": "/images/articles/source/images/sandbox-skills/skills.png",
       "sha256": "6ad86d972c7438fbef898d9d339f3f17a30277bf1626e8a9dd11b0b46583c817",
       "bytes": 362173
+    },
+    {
+      "path": "/images/articles/telegram-sandbox-qr.png",
+      "sha256": "8b4ead3dacb357fd28b650d2a6761b265effdeecc8e6a3f1f9feed10122da04b",
+      "bytes": 5583
+    },
+    {
+      "path": "/images/articles/telegram-sandbox.jpg",
+      "sha256": "0dab8f9c9c6846f8ac3b8848d118bca3528868254a7249d0d26772175eabae84",
+      "bytes": 349163
+    },
+    {
+      "path": "/images/articles/wecom-customer-service-qr.png",
+      "sha256": "3d2c2be11af63528a1b09871157ba58b4f3b64d7ddcbdf4420b99ecd41984c3b",
+      "bytes": 202053
+    },
+    {
+      "path": "/images/articles/wecom-customer-service.jpg",
+      "sha256": "2b94049e6265ffb1ac67cde661c596780d7fcfaeae46517ebd82f4dbfe3a1e19",
+      "bytes": 310163
+    },
+    {
+      "path": "/images/articles/wecom-knowledge-base-config.png",
+      "sha256": "0e4b12cec9bccf7c8a4958e9a53d312707ad818ac4376b707a20e8793af8f8be",
+      "bytes": 112097
     }
   ]
 }
diff --git a/docs/scripts/sync-blog-articles.mjs b/docs/scripts/sync-blog-articles.mjs
index a3c90d024..76c1f7f29 100755
--- a/docs/scripts/sync-blog-articles.mjs
+++ b/docs/scripts/sync-blog-articles.mjs
@@ -231,6 +231,8 @@ async function renderArticle(sourceRoot, post, locale, fallback, assetRoot, imag
   const config = LANGUAGE_CONFIG[locale];
   const title = locale === "ja" && fallback ? (JAPANESE_TITLES[post.slug] || post.data.title) : post.data.title;
   let body = post.body;
+  // Root-relative Blog links belong to the landing site, not the embedded docs.
+  body = body.replace(/(\]\()\/(en|zh|ja)\/blog\//g, "$1https://langbot.app/$2/blog/");
   const refs = imageReferences(body);
   for (const ref of refs) if (!imageCache.has(ref.url)) imageCache.set(ref.url, await materializeImage(sourceRoot, ref.url, assetRoot));
   if (post.data.cover?.image && !imageCache.has(post.data.cover.image)) imageCache.set(post.data.cover.image, await materializeImage(sourceRoot, post.data.cover.image, assetRoot));
diff --git a/docs/tests/agent-static-build.test.mjs b/docs/tests/agent-static-build.test.mjs
index a03bf5ee1..c477e5971 100644
--- a/docs/tests/agent-static-build.test.mjs
+++ b/docs/tests/agent-static-build.test.mjs
@@ -23,7 +23,8 @@ function links(node, result = []) {
 }
 
 test("every sitemap page has meaningful canonical Markdown and HTML discovery", async () => {
-  assert.equal(urls.length, 509 + zhOnlyArticleRoutes.length);
+  const manifest = JSON.parse(await readFile(path.join(root, "scripts/blog-articles-manifest.json"), "utf8"));
+  assert.equal(urls.length, 473 + zhOnlyArticleRoutes.length + manifest.articles.length);
   for (const url of urls) {
     const route = relative(url);
     const text = await read(`${route}.md`);
diff --git a/docs/tests/blog-articles.test.mjs b/docs/tests/blog-articles.test.mjs
index 3fe853c4a..1b5f68cda 100644
--- a/docs/tests/blog-articles.test.mjs
+++ b/docs/tests/blog-articles.test.mjs
@@ -58,6 +58,7 @@ test("generated articles use only non-empty Wiki-local images", () => {
     const file = path.join(root, `${article.page}.mdx`);
     assert.ok(existsSync(file), `missing ${article.page}.mdx`);
     const content = readFileSync(file, "utf8");
+    assert.doesNotMatch(content, /\]\(\/(?:en|zh|ja)\/blog\//, `${article.page} has a Blog link relative to the docs mount`);
     const imageUrls = [...content.matchAll(/!\[[^\]]*\]\(([^)]+)\)/g)].map((match) => match[1]);
     for (const imageUrl of imageUrls) {
       assert.match(imageUrl, /^\/images\/articles\//, `${article.page} hotlinks ${imageUrl}`);
diff --git a/docs/tests/locale-static-build.test.mjs b/docs/tests/locale-static-build.test.mjs
index 91ebc66db..6ececb51b 100644
--- a/docs/tests/locale-static-build.test.mjs
+++ b/docs/tests/locale-static-build.test.mjs
@@ -35,6 +35,7 @@ test("every localized HTML payload includes the preference client provider", asy
     }
   }
   for (const locale of ["en", "zh", "ja"]) await visit(path.join(root, locale));
-  // 509 canonical content pages plus one framework 404 page per locale.
-  assert.equal(count, 515);
+  // Every sitemap page plus one framework 404 page per locale.
+  const sitemap = await readFile(path.join(root, "sitemap.xml"), "utf8");
+  assert.equal(count, [...sitemap.matchAll(/<loc>/g)].length + 3);
 });